DAVID C. BROGAN
Visiting Assistant Professor
|Computer Graphics / Animation|
|Autonomous agents can
be realistically depicted through computer graphics. Animated
characters are needed to play the role of teachers or guides, teammates
or competitors, or just to provide a source of interesting motion in
virtual environments. For a virtual environment to be compelling, the
characters must have a wide variety of complex and interesting behaviors
and must be responsive to the actions of the user. The difficulty of
constructing such synthetic characters currently hinders the development
of these environments, particularly when realism is required. My
animation research draws from rigid body simulation, artificial
intelligence, and machine learning. I apply modeling and control
methods obtained from these fields to cutting-edge human modeling
applications. I study the biomechanics and cognitive strategies humans
use to accomplish locomotion and to plan coordinated movements in
complex environments. The products of my research are interactive
autonomous agents for immersive environments, training applications, and
scientific applications. Using this technology, we have analyzed
cerebral palsy gaits, pedestrian safety during building evacuations,
and learned group coordination and competition behaviors from
|Brogan, D. and Johnson, N. Realistic Human Walking Paths. In Proceedings of Computer Animation and Social Agents (CASA), pp. 94–101, 2003.|
|Brogan, D. and Hodgins, J. Simulation Level of Detail for Multiagent Control. In Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems (AAMAS), pp. 199-206, 2002.|
|Ventura, D. and Brogan, D.
Digital Storytelling with DINAH: Dynamic, Interactive,
Narrative Authoring Heuristic. In
Proceedings of the International Workshop on Entertainment Computing (IWEC),
pp. 91-99, 2002.
|Hodgins, J., Wooten, W., Brogan, D., O'Brien, J. Animating Human Athletics. In Proceedings of SIGGRAPH 1995, pp 71-78, 1995.|
|Brogan, D. C., Metoyer, R. A., and Hodgins, J. K., Dynamically Simulated Characters in Virtual Environments. IEEE Computer Graphics and Applications. September/October 1998, Volume 15 Number 5, p. 58-69. Also Animation Sketch in SIGGRAPH 1997.|
| Brogan, D.C., Hodgins, J.
Group Behaviors for Systems with Significant Dynamics
. Autonomous Robots
Brogan, D. and Hodgins, J. Group Behaviors for Systems with Significant Dynamics. In Proceedings of the 1995 IEEE/RSJ International Conference on Intelligent Robots and Systems, Vol. 3, pp 528-534, 1995.
Hodgins, J. and Brogan, D. Robot Herds: Group Behaviors for Systems with Significant Dynamics. In Proceedings of Artificial Life IV, 319-324, 1994.
|Brogan, D., Granata, K., and Sheth, P.
Spacetime Constraints for Biomechanical Movements.
In Proceedings of Applied Modeling and Simulation, 2002.
Granata, K., Brogan, D., and Sheth, P. Stable Forward Dynamic Simulation of Bipedal Gait Using Space-Time Analysis. In American Society of Biomechanics Proceedings of IV World Congress of Biomechanics, 2002.
|Autonomous Agents / Swarms|
Computer scientists began investigating artificial life
and the emergent properties of agent-based simulations in the mid 1980s.
Inspired by the robust decentralized and redundant control methodologies
observed in biological systems, they explored the potential for replacing
centralized algorithms with distributed sensing and decision making. The
alluring simplicity of creating agents that could be duplicated indefinitely was
counterbalanced by the myriad challenges of determining how many agents were
necessary and what each agent needed to compute, sense, and communicate to
produce the desired group behaviors. What arises is a high-dimensional,
non-linear system of differential equations that is difficult to understand and
debug due to the complexity of interactions among the agents. Data-driven
technologies and multi-resolution modeling can simplify these complex systems
and emphasize specific features of system behavior. Our mutiagent modeling and
simulation techniques are demonstrated with autonomous crowds that escape
burning buildings and soccer players that improve their game by observing
|Ratner, O. and Brogan, D. Simulating Crowds with Balance Dynamics. Poster presented at ACM SIGGRAPH 2005.|
|White, C. and Brogan, D. Reinforcement Learning in Simulated Soccer with Kohonen Networks. Presented as a poster at 9th INFORMS Computing Society Conference, 2005.|
|Loitiere, Y., Brogan, D, and Reynolds, P. Simulation Coercion Applied to Multiagent DDDAS. Workshop on DDDAS in Proceedings of the International Conference on Computational Science (ICCS), pp. 789-796, 2004.|
|Brogan, D. and Loitiere, Y. Building Multiagent Behaviors from Observation. In Workshop on Intelligent Human Augmentation and Virtual Environments, 2002.|
|Brogan, D. and Loitiere, Y. Data-Driven Generation of Simulated Soccer Behaviors. In Proceedings of the International Joint Conference on Autonomous Agents and Multiagent Systems (AAMAS), pp. 1391-1392, 2002.|
|Browne, K., McCune, J., Trost, A., Evans, D., and Brogan, D. Behavior Combination and Swarm Programming. In RoboCup International Symposium 2001, Lecture Notes in Artificial Intelligence, Springer-Verlag, pp. 499-502, 2001.|
As scientists and engineers increase their use of
simulations and computational tools, previous and current research by computer
scientists becomes more relevant. Simulations are becoming larger, remaining in
use longer, and are serving bigger communities. They can be overwhelmingly
complex. Much of the complexity stems from the interface between the analytical
scientific models under study and the numerical methods that implement those
models. A scientist must prescribe such implementation details as temporal and
spatial resolution, the numerical precision of variables, and the operating
assumptions (ignore wind resistance in a falling-body problem). Similarly, the
scientific model must adapt to the numerical accuracy of the implemented
differential equation integration algorithms and the computational limitations
of the underlying computer hardware. My research reveals that what scientists
and engineers need most are tools to articulate simulations’ assumptions,
limitations, and capabilities so they may more easily be verified, validated,
and integrated in novel contexts. These simulation tools cannot simply
take the form of a Matlab extension or an off-line visualization program; they
must be tightly integrated into the end-to-end lifespan of a simulation from its
genesis as the core research in one scientist’s lab to its use years later as a
small piece of another scientist’s system. Using these tools, a simulation
designer will provide hints (growth paths), provisions (code expansions), and
insights (sensitivity analyses) that guide later users and developers regarding
possible later simulation adaptations when novel requirements arise.
|Spiegel, M., Reynolds, P., and Brogan, D. A Case Study of Context Assumptions for Simulation Composability and Reusability. To appear in the Proceedings of the Winter Simulation Conference (WSC), 2005.|
Brogan, D., Reynolds, P., Bartholet, R., Carnahan, J., and Loitiere, Y. Semi-Automated Simulation Transformation for DDDAS. Workshop on DDDAS in Proceedings of the International Conference on Computational Science, pp. 721-728, 2005.
Bartholet, R. Brogan, D., and Reynolds, P. The Computational Complexity of Component Selection in Simulation Reuse. To appear as a poster and printed as full article in Proceedings of the Winter Simulation Conference (WSC), 2005.
Fox, M., Brogan, D., and Reynolds, P. Approximating Component Selection. In Proceedings of the Winter Simulation Conference (WSC), pp. 429-435, 2004.
Bartholet, R., Reynolds, P., and Brogan, D. Semantics, Scope, or Scale: Simulation Composability Versus Component-Based Software Design. In Proceedings of the Simulation Interoperability Workshop (SIW), 2004.
Carnahan, P., Reynolds, P, and Brogan, D.
Visualizing Coercible Simulations. In Proceedings of the
Winter Simulation Conference (WSC), pp. 411-420, 2004.
Carnahan, J., Brogan, D. and Reynolds, P. Simulation-Specific Characteristics and Software Reuse. To appear as a poster and printed as full article in Proceedings of the Winter Simulation Conference (WSC), 2005.
Carnahan, J., Reynolds, P., and Brogan, D. Language Support for Identifying Flexible Points in Coercible Simulations. In Proceedings of the Simulation Interoperability Workshop (SIW), 2004.
|Waziruddin, S., Brogan, D., and Reynolds, P. Selecting Optimization Techniques in Support of Simulation Transformation. In Proceedings of the Simulation Interoperability Workshop (SIW), 2004.|
Carnahan, J, Reynolds, P., and Brogan, D. An Experiment in Simulation Coercion. In Proceedings of the Interservice / Industry Training, Simulation, and Education Conference (I/ITSEC), 2003.
Giordano, J., Reynolds, P., and Brogan, D. Synthetic Worlds, Authentic Bounds: Exploring the Characteristics and Constraints of Human Behavior Representation. In Proceedings of the Winter Simulation Conference (WSC), pp. 912-921, 2004.
|Waziruddin, S., Brogan, D., and Reynolds, P. The Process for Coercing Simulations. In Proceedings of the Simulation Interoperability Workshop (SIW), 2003.|
Department of Computer Science
School of Engineering & Applied Science
University of Virginia
151 Engineer's Way, P.O. Box 400740
Charlottesville, VA 22904-4740
office: 217 Olsson Hall
phone: (434) 982-2211